“Class A” is a marketing buzz word in guitar and hi-fi tube audio these days. Unfortunately most musicians don’t know the origins of the term, and incorrectly assume it relates to the quality of the design, construction, or parts. Class A is an engineering term describing how an amplifier is biased. Class A are the simplest amplifiers, in terms of required components. Other amp classes (AB, B, C, D, etc...) add additional parts and complexity to gain higher efficiency.
There are many alphabetic amplifier classes, but for analog audio we really need only consider class A and AB. Also, this discussion really only applies to power amps; audio preamps are always class A. Related articles are Single Ended vs Push Pull and Measuring Power.
Class A is the least efficient amplifier class, with a maximum theoretical efficiency of only 50%. This means that for every watt the amplifier consumes from the power supply, it delivers half a watt into the load (a speaker, for instance). Put another way, for every watt delivered to the load the amplifier requires double that, two watts, from the power supply. The power not delivered (the other 50%) is dissipated in the tube as heat.
A class A amplifier consumes the same amount of power (from the power supply) whether there is an input signal present or not. For example consider a 5W amplifier; this amp can deliver up to 5 watts to the load, but will consume 10 watts (at least), even when there is no signal. When there is no input signal no power is delivered to the load; all of it is dissipated as heat (0% efficiency).
The inefficiency of class A generates heat and requires a large, heavy duty, expensive power supply to provide the necessary current. Heat is one of the enemies of electronics, and manufacturers are always working to reduce the size, weight, and cost of their products, so most audio amplifiers are not designed to operate in class A. However, class A is the most linear amplification class: in class A push-pull each amplifying device (such as a tube, transistor, IC, etc) amplifies the entire signal. It is also the only amplifier class which can operate in both single ended and push-pull modes.
Class A waveforms : Each side of the push-pull circuit amplifies the whole signal
More efficient that class A, class B amplifiers have a maximum theoretical efficiency around 75%. However, unlike class A, class B requires a pair of amplifying devices (tubes, transistors, ICs, etc) and must operate in push-pull (also called “differential” mode). Class B amplifiers split the input signal into two halves (with opposite phases - see Single Ended vs Push Pull; one device amplifies the top half of the signal and the other amplifies the bottom half. The two halves are then recombined into one signal at the output (by taking the difference between the two halves).
Class B waveforms : Each side of the push-pull circuit amplifies exactly half of the signal
Class B amplifiers look great on paper but in practice the amplifying devices will always have some mismatch, resulting in overlap or dead time when one device is turning off and the other is turning on. These events do not occur at exactly the same time, causing glitches in the output known as “crossover” or “notch” distortion. This sounds terrible and prevents class B amplifiers from being used for audio. Also, because they require at least two tubes, class B amplifiers cannot be single ended.
Class B vs Class A
Class C is the most efficient of the analog amplifier classes, with maximum theoretical efficiency approaching 90%. However they are extremely nonlinear and are of little use in audio applications due to significant crossover distortion.
Class C waveforms : Each side of the push-pull circuit amplifies less than half the signal
Class C vs Class A
Class AB is a loose definition for any amplifier operating between class A and class B. Engineers found they could reduce the crossover distortion by applying a small positive bias to both tubes (bias shown in red on the graphs below). This means that the amplifier will dissipate some power when there is no input signal present, but it is nowhere near the power dissipated by a class A amplifier, so the heat generated by class AB is lower as well.
Class AB waveforms : Each side of the push-pull circuit amplifies more than half the signal
Class AB is an engineering compromise between the higher efficiency of class B and the lower distortion of class A. Crossover distortion still occurs in class AB amplifiers, but only at larger signal levels, depending on how high the bias is set. If the bias were raised so high that crossover distortion never occurred the amplifier would be class A; if it were lowered to zero, where crossover distortion always occurs, it would be class B.
The power supply for a class AB amplifier can be built smaller and cheaper than for a class A amplifier since the required bias current is lower. However, because the power supply is less robust than in a class A amplifier, the bias for a class AB amplifier must be carefully adjusted. If the bias is too high the power supply will not operate as designed, if too low the amp will not sound good (too much crossover distortion). Class AB amps usually call for matched power tubes, also an attempt to reduce crossover distortion, but even matched tubes will perform differently over time. For this reason class AB tube amps must typically have their bias checked often, not only when installing new tubes, but also as the tubes age and drift.
In electrical terms, a circuit’s bias is where the current will settle when there is no input signal. This point is also referred to as “idle” or “quiescent”. It may help to think of a tube as a spring-loaded leaky water faucet, with some amount of water flowing through it. You can hold your hand on the faucet to turn it off or further on, but when you release your hand (the input signal) the faucet springs back to the set bias point. A class B faucet would turn itself all the way off when not in use and would not leak at all. A class A faucet would run at half it’s maximum flow when not in use, and a class AB faucet would drip or slowly trickle.
All amplifier stages have a bias point, though most are not adjustable. For reasons mentioned above, class AB tube power amplifiers typically have an adjustable bias. An amplifier’s bias point determines the class of operation and affects performance metrics like gain, distortion/linearity, noise, power output, efficiency, etc.
The Vox AC-30 (AC stands for Amplifier Combination) is one of the best known “class A” amplifiers. It is reported to deliver 30W from four EL84 tubes. As you may recall, the maximum theoretical efficiency of a class A amplifier is 50%. From the EL84 data sheet, the maximum power dissipation (Wa) of the EL84 tube is 12W.
4 tubes * 12W * 50% = 24W RMS in class A (output power is measured in watts RMS)
Of course these numbers are theoretical and optimistic; in practice 20 watts is about the best we can do with four EL84s in class A. So the Vox AC-30 is either not class A, or it is not 30 watts; it can’t be both.
There are quite a few amps claiming to be “class A”. It’s easy to verify if the claims are true with some simple arithmetic. The maximum safe power dissipation of a given tube can be found on that tube’s data sheet (usually called “plate dissipation” or "Wa"); search on Google for “EL84 data sheet”, etc.
In summary, if you need high power (more than 30W) class AB is really your only choice. If you don’t need high power (under 30W) and want a smaller, lighter, or cheaper amplifier you should also go with class AB. If you don’t need high power and want the absolute best sound go with class A. If you want low power (5W and below) go with a single ended amplifier; all of which are class A.